1 / 25

Yuanlu Xu Advisor: Prof. Liang Lin merayxu@gmail

Yuanlu Xu Advisor: Prof. Liang Lin merayxu@gmail.com. Person Re-identification by Matching Compositional Template with Cluster Sampling. Problem. Person Re-identification. Identifying The Same Person Under Different Cameras. Basic Assumption :

hinda
Télécharger la présentation

Yuanlu Xu Advisor: Prof. Liang Lin merayxu@gmail

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Yuanlu Xu Advisor: Prof. Liang Lin merayxu@gmail.com Person Re-identification by Matching Compositional Template with Cluster Sampling

  2. Problem Person Re-identification Identifying The Same Person Under Different Cameras Basic Assumption: Face is unreliable due to view, low resolution and noises. People's clothes should remain consistent.

  3. Difficulty Large Intra-class Variations Pose/View Variation Illumination Change Occlusion

  4. Problem Multiple Setting Query Person Scene S vs. S M vs. S Search

  5. Representation Multiple-Instance Compositional Template (MICT) • Body into 6 parts, limbs further into 2 symmetric parts. • Leaf nodes contain multiple instances. • Contextual relationsbetween parts: • kinematics • symmetry.

  6. Problem Formulation Matching-based Formulation … • Given the template, the problem is formulated as • Selecting an instance for each part. • Finding the matched part in target.

  7. Problem Formulation • Candidacy Graph: • Vertices – possible matching pairs

  8. Problem Formulation • Solving the problem: • Labeling vertices in the graph (selecting matching pairs) • NP hard – incorporating graph edges

  9. Problem Formulation • Compatible Edges: • Encouraging matching pairs to activate together in matching • Defined by contextual constraints

  10. Problem Formulation • Competitive Edges: • Depressing conflicting matching pairs being selected at the same time • Defined by matching constraints

  11. Inference Using Cluster Sampling [1]for inference: Sampling edges in candidacy graph to generate clusters. Randomly selecting/deselecting the clusters. Decide whether to accept the new state. [1] J. Porway et al.,“C4: Exploring multiple solutions in graphical models by cluster sampling”, TPAMI 2011.

  12. Dataset VIPeR Dataset: 1. Classic ReIDdataset 2. Well-segmented people, limited pose/view 3. Heavy illumination changes,lack occlusion D. Gray et al., "Viewpoint Invariant Pedestrian Recognition with an Ensemble of Localized Features”, ECCV 2008.

  13. Dataset Query Instance Video Shot Target Individual EPFL Dataset: 1. Cross-camera tracking dataset 2. Few people, shot scene provided, various pose/view 3. Little illumination changes, limited occlusions F. Fleuret et al., "Multiple Object Tracking using K-Shortest Paths Optimization”, TPAMI 2011.

  14. Dataset Query Instance Video Shot Target People CAMPUS-Human Dataset: 1. Camera and annotate by us 2. Many people, shot scene provided, various pose/view 3. Limited illumination changes, heavy occlusions

  15. Result Setting 1: Re-identify peoplein segmented images, i.e. targets already localized.

  16. Result Setting 2: Re-identify peoplefrom scene shots without provided segmentations.

  17. Result Component Analysis Evaluating feature and constraints effectiveness

  18. Conclusion A solution for a new surveillance problem. A person-based model, a graph-matching-based formulation, a more complete database for evaluation. Exploring robust and flexible person models [1], efficient search method [2] in future. [1] J. B. Rothrock et al., “Integrating Grammar and Segmentation for Human Pose Estimation”, CVPR 2013. [2] J. Uijlings et al., “Selective Search for Object Recognition”, IJCV 2013.

  19. Published Papers Yuanlu Xu, Liang Lin, Wei-Shi Zheng, Xiaobai Liu. “Human Re-identification by Matching Compositional Template with Cluster Sampling”. ICCV 2013. Liang Lin, Yuanlu Xu, Xiaodan Liang, Jian-Huang Lai. “Complex Background Subtraction by Pursuing Dynamic Spatio-temporal Manifolds”. IEEE TIP 2014, under revision. Yuanlu Xu, Bingpeng Ma, Rui Huang, Liang Lin. “Person Search in a Scene by Jointly Modeling People Commonness and Person Uniqueness”. ACMMM 2014, submitted.

  20. QUESTIONS?

  21. Cluster Sampling Generating a composite cluster 1. Given a candidacy graph and the current matching state , we first separate graph edges into two sets: set of inconsistent edges and set of consistent edges in the other two cases. 2. Next we introduce a boolean variable to indicate an edge is being turned on or turned off. We turn off inconsistent edges deterministically and turn on every consistent edge with its edge probability .

  22. Cluster Sampling Generating a composite cluster 3. Afterwards, we regard candidates connected by ”on” positive edges as a cluster and collect clusters connected by ”on” negative edges to generate a composite cluster .

  23. Composite Cluster Sampling Using Metropolis-Hastings method to achieve a reversible transition between two states and , the acceptance rate of the transition is defined as posterior ratio state transition probability ratio

  24. Composite Cluster Sampling The state transition probability ratio is computed by edges being turned off around ,

  25. Composite Cluster Sampling Inference Algorithm

More Related